Crossover distortion and quasi-complementary class AB

Crossover distortion

In a class B push-pull amplifier, each transistor conducts only when its base-emitter voltage exceeds about 0.6 to 0.7 V. For a small input, between about −0.7 V and +0.7 V, neither transistor conducts, so the output stays at zero near the zero-crossing. The output waveform is therefore flattened (a dead zone) at each crossing between the positive and negative half-cycles. This harmonic distortion is called crossover distortion. It is most serious for small signals.

Quasi-complementary symmetry class AB amplifier

When complementary power transistors (a matched NPN and PNP) are not available, both output transistors are NPN. The upper half is an NPN Darlington. The lower half uses a PNP driver and an NPN power device in a feedback-pair (Sziklai) arrangement, which behaves like a single PNP transistor.

  +VCC ---+------------------+------------------+
          |                  |                  |
        [R1]                 C                  C
          |                  |                  |
          +--------------B  Q1 (NPN driver)     |
          |                  E --------------B  Q2 (NPN power)
         D1                                     E
         D2                                     |
 Vi o-||--+                                    [Re]
         D3                                     |
          |                                     +--------o Vo
          |                                     |        |
          |                                    [Re]     [RL]
          |                                     |        |
          |                  +------------------+       GND
          |                  E                  C
          +--------------B  Q3 (PNP driver)     |
          |                  C --------------B  Q4 (NPN power)
        [R2]                                    E
          |                                     |
  -VCC ---+-------------------------------------+

(Upper half: NPN Darlington Q1–Q2. Lower half: the PNP driver Q3 drives the base of the NPN power transistor Q4; Q4's collector is joined to Q3's emitter, forming the feedback pair, and Q4's emitter goes to −VCC. Exact wiring varies slightly between textbooks.)

How class AB removes crossover distortion

1. The diodes D1–D3 (or a VBE multiplier) between the two driver bases drop about 2 V, one diode drop for each base-emitter junction in the path (two in the upper Darlington, one for Q3). This is just enough to forward bias the output stage.
2. At zero input both output halves carry a small quiescent current (class AB). The conduction angle of each is slightly more than 180°.
3. When the input goes positive, the upper transistor increases its current while the lower one smoothly reduces its current to zero (and vice versa). The two currents add through the crossover with no gap.
4. The diodes are mounted close to the output transistors, so the bias tracks temperature and prevents thermal runaway. The small Re resistors add stability.

The output is then a smooth sine wave. The cost is a small idle current and an efficiency a little lower than that of ideal class B.